Research On Motion Object Detection And Tracking Based On Complex Background

Currently, Moving object detection and tracking in image sequence is one of thecritical issues in the field of computer vision for its widely applications in the fields ofmilitary guidance, intelligent traffic, safety monitoring and others. Because of theinfluences from the factors including illumination changes, water wave shaking, plantswing and so on, research on moving object detection and tracking algorithm in thecomplex background, has become very urgent.Image preprocessing, moving objects detection and objects tracking are mainlyincluded in this paper. On the basis of the analysis and summary of existing algorithmswhich are commonly used in moving object detection and tracking, the improvedbackground subtraction based on GMM is presented. The way can be used to detectmoving object based on complex background. Kalman Filter can effectively forecast thestate of moving target, so it can be used to reduce the time of matching operations which iscost by Mean Shift algorithm. Using improved Mean Shift algorithm and movablechannel-filter for monitoring pedestrians in different areas, we got excellent detection andtracking effect. As the same time, we use special background subtraction for defendingtagging along, and it also runs well.